Copyright ã Author(s) 2001 * [email protected]This is a revised version of the paper originally prepared for the UNU/WIDER development conference on Debt Relief, Helsinki, 17-18 August 2001. UNU/WIDER gratefully acknowledges the financial contribution from the governments of Denmark, Finland and Norway to the 2000-2001 Research Programme. Discussion Paper No. 2001/100 Does the HIPC Initiative Achieve its Goal of Debt Sustainability? Bernhard G. Gunter * September 2001 Abstract This paper examines the question if the Heavily Indebted Poor Country (HIPC) Initiative provides a good basis for the HIPCs to exit from repeated debt rescheduling. Building on other reviews of the HIPC Initiative, the paper begins with a short summary of some key problems of the HIPC Initiative. It then reviews critically the growth assumptions of HIPC debt sustainability analyses, whereby the paper examines the changes in (i) public and private capital flows before and after the adoption of the HIPC Initiative, (ii) investment and savings rates, and (iii) sectoral transformations. The last analytical part explores the appropriateness of the HIPC debt sustainability indicators. Before summarizing the main results, the paper makes some suggestions on possible modifications in the HIPC framework that are more likely to provide debt sustainability than the current framework. Keywords: debt sustainability, structural change, growth JEL classification: F34, O11, F35
This is a revised version of the paper originally prepared for the UNU/WIDER development conferenceon Debt Relief, Helsinki, 17-18 August 2001.
UNU/WIDER gratefully acknowledges the financial contribution from the governments of Denmark,Finland and Norway to the 2000-2001 Research Programme.
Discussion Paper No. 2001/100
Does the HIPC Initiative Achieveits Goal of Debt Sustainability?
Bernhard G. Gunter *
September 2001
Abstract
This paper examines the question if the Heavily Indebted Poor Country (HIPC)Initiative provides a good basis for the HIPCs to exit from repeated debt rescheduling.Building on other reviews of the HIPC Initiative, the paper begins with a short summaryof some key problems of the HIPC Initiative. It then reviews critically the growthassumptions of HIPC debt sustainability analyses, whereby the paper examines thechanges in (i) public and private capital flows before and after the adoption of the HIPCInitiative, (ii) investment and savings rates, and (iii) sectoral transformations. The lastanalytical part explores the appropriateness of the HIPC debt sustainability indicators.Before summarizing the main results, the paper makes some suggestions on possiblemodifications in the HIPC framework that are more likely to provide debt sustainabilitythan the current framework.
UNU World Institute for Development Economics Research (UNU/WIDER)was established by the United Nations University as its first research andtraining centre and started work in Helsinki, Finland in 1985. The purpose ofthe Institute is to undertake applied research and policy analysis on structuralchanges affecting the developing and transitional economies, to provide aforum for the advocacy of policies leading to robust, equitable andenvironmentally sustainable growth, and to promote capacity strengtheningand training in the field of economic and social policy making. Its work iscarried out by staff researchers and visiting scholars in Helsinki and throughnetworks of collaborating scholars and institutions around the world.
UNU World Institute for Development Economics Research (UNU/WIDER)Katajanokanlaituri 6 B, 00160 Helsinki, Finland
Camera-ready typescript prepared by Liisa Roponen at UNU/WIDERPrinted at UNU/WIDER, Helsinki
The views expressed in this publication are those of the author(s). Publication does not implyendorsement by the Institute or the United Nations University, nor by the programme/project sponsors, ofany of the views expressed.
Though the author is currently a consultant in the World Bank’s Africa Region, and wasan economist in the World Bank’s HIPC unit from 1998-2000, this paper is not relatedto either position and the views expressed here are his own. The views should not beassociated to the World Bank, its executive directors, or the countries they represent.Some valuable comments from conference participants are acknowledged. If nototherwise stated, the data is taken from the World Bank’s Global Development Finance2001.
1
1 Introduction
Today, it is a well-known fact that unsustainable debt has a negative impact oninvestment, growth, and development; the so-called debt-overhang effect. The empiricalevidence provided since the early 1990s for a debt overhang in many of the poorest andhighly indebted countries has led—though with considerable delay—to the adoption ofthe Heavily Indebted Poor Country (HIPC) Initiative in fall 1996.1 Compared to earlierdecades of bilateral debt rescheduling, the original framework of the HIPC Initiativewas a major break-through, mainly due to the HIPC Initiative’s key goal to reduce thedebt of the poorest countries to a level that would allow them to permanently exit theprocess of repeated debt rescheduling.
However, three years after launching the HIPC Initiative, it was clear that the originalHIPC framework was not a sufficient solution for many poor countries to reach debtsustainability. Largely due to public pressure, IMF and World Bank agreed inSeptember 1999 to enhance the HIPC framework. The enhancements provide broader,deeper and faster debt relief mainly through (i) a lowering of the ratios considered toprovide debt sustainability (together with a lowering of the minimum thresholds toqualify for the openness/fiscal criteria), (ii) replacing the principally fixed three-yearperiod between decision and completion points by the concept of a floating completionpoint, and (iii) the provision of interim relief from some creditors between the decisionpoint and the completion point. Another key enhancement was to link HIPC debt reliefto the preparation of country-owned poverty reduction strategies.
Nevertheless, there remain many problems with the enhanced HIPC Initiative. First ofall, evidence is once again mounting that even the enhanced HIPC framework does notprovide long-term debt sustainability for many of the poorest countries, mainly because(i) its growth assumptions are considered too optimistic, (ii) its debt sustainabilityanalysis inappropriate, and (iii) its country selection too narrow. For example, in spring2000, the United States General Accounting Office (GAO 2000) concluded that theHIPC Initiative might not provide a lasting exit from debt problems, unless strong andsustained economic growth is achieved. The report cautions that the growth assumptionsused by IMF and World Bank staff for the country-specific debt sustainability analyses(DSAs) may be overly optimistic.2
Recognizing the possibility that the HIPC Initiative may not achieve debt sustainability,the IMF and World Bank have recently issued a paper on the challenge of maintaininglong-term debt sustainability.3 The paper emphasizes the importance of establishing anenvironment conducive to growth and poverty reduction, particularly in the areas ofmacroeconomic policies, structural reforms, public sector management, governance andsocial inclusion. It also notes that HIPCs are typically dependent upon a narrow exportbase, which makes them vulnerable to externally induced shocks. It examines thesensitivity of long-term debt sustainability to possible shortfalls in export revenues and
1 The theoretical rational for the negative impact of a debt overhang on investment and growth havebeen provided by the seminal contributions by Sachs (1989) and Krugman (1988); for a listing of theearly empirical studies, see Gunter (forthcoming).
2 For example, IMF and World Bank assume that export earnings will grow in excess of 9 per centevery year for 20 years in four of the seven HIPCs the GAO analysed (GAO 2000: 9).
3 IMF and World Bank (2001).
2
less concessional financing than assumed in the DSAs, yet, the paper does not assess thelikelihood of these and other factors influencing growth prospects. Without addressingthe broad critique of possibly overly optimistic growth rates,4 the paper concludes thatthe HIPC Initiative provides a good basis for HIPCs to exit from future debtrescheduling.
As reviewed in more detail in Gunter (forthcoming), other major critiques to theenhanced HIPC Initiative can be grouped into overall problems with the HIPCframework and specific problems related to HIPC debt relief. Overall problems of theHIPC framework are that (i) developing countries’ suggestions have not been takenserious enough, (ii) the initiative’s burden sharing is unrelated to economic power,(iii) the HIPC Initiative is confronted with major financing problems, some of whichhave been pushed to deal with in the future, (iv) the anticipation of HIPC eligibility islikely to defer traditional development assistance, and (v) discounts rates are usedinappropriately and inconsistently. Given that all of these problems have beenmentioned in one way or the other in the extensive HIPC review undertaken before theadoption of the enhanced framework, a broader discussion of the various suggestionscould have avoided many of the current problems. More specific, but equally crucialproblems are that HIPC debt relief is (i) not calculated based on a country’s need forsustainable development, (ii) likely to be deducted from traditional developmentassistance, (iii) unnecessarily delayed by the adoption and implementation of povertyreduction strategies, and (iv) partly delivered through debt rescheduling.
2 Growth assumptions of HIPC DSAs
Table 1 shows the actual growth rates in real GDP and the projected real growth ratesassumed in HIPC DSAs of the 22 countries that had reached the enhanced decisionpoint by December 2000. Based on the experience of the 1990s and without analysingthe growth projections further, some projections seem realistic (i.e., for Uganda andMozambique), however, most projections seem highly unrealistic. Considering worldhistory, any long-term real GDP growth rate of more than 6 per cent is highlyexceptional. It seems unlikely that Mauritania, Guinea-Bissau, Madagascar, andRwanda will repeat what has been known as the East Asian miracle. The averagegrowth rates for 2000-10, assumed for the first 22 enhanced decisions-point countries,are 5.5 per cent for real GDP and 8.6 per cent for exports (expressed in nominal USdollars, ranging from 4.4 per cent for Guyana to 13.7 per cent for Rwanda).
Too optimistic growth rates affect the HIPC framework’s debt sustainability in twoways: first, they imply too optimistic growth rates of a country’s exports, and second,they underestimate a country’s future financing needs. Overestimations of exports(which are in the denominator of the ratio) and underestimations of future financingneeds/new debt (which are in the nominator of the ratio) result in highly unrealistic lowfuture debt-to-export ratios, which then indicate unrealistic long-term debtsustainability. As the GAO (2000: 15) report points out, if Tanzania’s exports grow atan annual 6.5 per cent (instead of the 9 per cent projected by the IMF and World Bank),
4 For example, even the World Bank (2001a: 102) has cautioned that the projected growth rates may notbe realistic.
3
Tanzania’s debt-to-export ratio could be more than twice of what the IMF’s and WorldBank’s forecast shows for the projection period.
At a more general level, the impact of various levels of export growth rates on long-term debt-to-export ratios are illustrated in Figure 1, showing the paths of NPV debt-to-export ratios of export growth rates of 5 per cent, 7 per cent, and 9 per cent. Note thatwe keep the NPV debt constant in all three cases. Thus, the reductions in NPV debt-to-export ratios are simply due to growth in exports. Comparing these NPV debt-to-exportratios with the projections as they can be found in most HIPC documents, it becomesclear that most of the projected NPV reductions are not due to HIPC debt relief, but dueto optimistic export growth rates.
Table 1Real GDP growth, 1990-99 and 2000-10
Real GDP growth,1990-99 average
GDP growth, 2000-10,HIPC DSA assumptions
Difference inpercentage points
Mauritania 4.3 7.3 3.0
Guinea-Bissau 0.3 7.0 6.7
Madagascar 1.8 6.2 4.4
Rwanda -1.6 6.1 7.7
Cameroon 1.2 6.0 4.8
Mozambique 6.3 5.9 -0.4
Honduras 3.2 5.9 2.7
Burkina Faso 3.6 5.9 2.3
Tanzania 3.1 5.9 2.8
Gambia, The 3.0 5.6 2.6
Uganda 6.7 5.6 -1.1
Nicaragua 2.6 5.6 3.0
Benin 4.3 5.5 1.2
Guinea 3.9 5.3 1.4
Bolivia 4.1 5.3 1.2
Zambia 1.0 5.2 4.2
Senegal 3.0 5.0 2.0
Mali 3.4 5.0 1.6
Malawi 4.0 4.4 0.4
Niger 2.4 4.4 2.0
Guyana 6.0 4.2 -1.8
Sao Tome & Principe -0.5 4.1 4.6
Source: IMF and World Bank (2001: Table 5)
4
Figure 1Evolution of NPV debt-to-export ratios
with zero NPV debt reduction
0
50
100
150
200
2000 2004 2008 2012 2016 2020
%
5% growth of exports7% growth of exports9% growth of exports
Today, the international community understands better than ever what the sources ofeconomic growth are: macroeconomic stability, increases in labour participation rates,investments in physical and human capital, increasing saving rates and financial marketdevelopment, openness to trade and investment, and a political system that provides forhuman rights and freedom, effective governance, and increasing democratization.5
Furthermore, there are also a couple of African-specific growth analyses that emphasizethe role of a hostile internal and external environment (ethnic diversity, lack of socialcapital, particularly dysfunctional government, limited trade openness, conflict and slowgrowth in neighbouring countries).6 Finally, there is a large literature that stresses thatsustainable growth can only be achieved if the economy undergoes a structuraltransformation, which results in export-led growth in manufactures.7 Based on thisbackground, we look at four aspects that can shed some light on future growth prospectsof HIPCs: (i) changes in capital inflows, (ii) changes in investment and savings ratios,(iii) an economy’s structural transformation, and (iv) implications of AIDS, climatechange, and armed conflict.
5 Until recently, the new growth literature was a collage of theoretical and empirical studies, many ofthem stressing the importance of one or a few sources of growth. Furthermore, there were someinfluential studies suggesting that most of the high growth experiences were due to a rapidaccumulation of labour and capital (either physical capital or human capital, or both). See Krugman(1994) and Young (1995). However, over the last few years, evidence has been mounting that theaccumulation of labour and capital do not explain the huge differences in growth experiences acrosscountries. Instead, attention has been shifting towards the residual representing total factorproductivity. For example, see Easterly and Levine (2000) and Senhadji (2000).
6 See Branson, Guerrero and Gunter (forthcoming) for a list of empirical growth analyses for Sub-Saharan Africa.
7 For example, see Sachset al. (1999: 12). For further references, as well as for an assessment of therole of capital accumulation and adjustment, see Berthélemy and Söderling (2001).
5
2.1 Capital flows before and after the HIPC Initiative
In this section, we look at various capital flows before and after the adoption of theHIPC Initiative for two reasons. First, certain increases in capital flows could be viewedas indicators for a country’s debt sustainability, and second, changes in the amount andcomposition of capital flows have direct implications on a country’s debt sustainability.Both factors are related to the fact that the growth prospects of HIPCs will continue todepend crucially on future foreign capital inflows. In this regards, most observers agreethat a necessary condition for the success of the HIPC Initiative is that debt relief shouldbe additional to the existing resource transfers.8 Previous analyses found little evidenceof additionality with regards to the resource transfers to HIPCs.9 Based on the latestdata generally available, we compare various capital flows during the three years beforeand after the adoption of the HIPC Initiative, i.e., we compare capital flows of 1994-96with those of 1997-99. We first look at private non-guaranteed capital flows, then atofficial capital flows, and finally at private but publicly guaranteed capital flows.
2.1.1 Changes in private non-guaranteed capital flows
The formal adoption of the HIPC Initiative in fall 1996 could have been a positivesignal for the future prospects of HIPCs, and given that private non-guaranteed debt isexcluded from HIPC debt relief, we could have expected an overall increase in privatenon-guaranteed capital inflows to HIPCs after the adoption of the HIPC Initiative, or atleast after a country reached its decision or completion point. However, the overallpicture in this regard is quite dim.
First, the total disbursement of non-guaranteed debt to the group of HIPCs decreasedfrom more than US$1.54 billion10 during 1994-96 to less than US$1.39 billion during1997-99. Excluding Bolivia, disbursements of commercial banks to the group of HIPCsdecreased from US$1.1 billion during 1994-96, to US$0.85 billion during 1997-99,even though disbursements of commercial banks to Sub-Saharan Africa increasedduring the same time from US$1.98 billion to US$3.04 billion. Second, while the groupof HIPCs received a marginal US$250 million in disbursements of private non-guaranteed bonds during 1994-96, no such bonds were issued during 1997-99. The onlypositive trend in private capital flows to the group of HIPCs is with regards to net flowsof foreign direct investment (FDI), which have increased by US$8 billion (from US$13billion during 1994-96, to US$21 billion during 1997-99), though about US$5 billion ofthis increase was due to sharp increases in FDI to Angola and Bolivia.
We now look at the changes in private capital inflows to the four HIPCs (Uganda,Bolivia, Burkina Faso, and Guyana) that have—due to their good track record—beenthe first countries reaching the decision point under the original framework of the HIPCInitiative in 1997. These four HIPCs have been outstanding performers in terms ofmacroeconomic policies and structural reforms for many years, which justified their
8 See UNCTAD (2000).
9 For example, see the Concluding Report of the November 2000 Bretton Woods CommitteeRoundtable Discussion on Reassessing Debt Relief (available at: www.brettonwoods.org), whichstates that roundtable participants, including Bank and Fund officials, were discouraged that at anearly stage, little evidence of additionality could be found.
10 We use the US definition, whereby one billion is thousand millions: 1.5 billion is 1,500 million.
6
original HIPC decision points in 1997. Uganda and Bolivia even reached theircompletion points under the original HIPC framework in 1998.
Though Uganda was unable to attract any private capital flows in terms of non-guaranteed loans or bonds, it was successful in attracting some foreign directinvestment, whereby the net inflows increased steadily from US$88 million in 1994 toover US$220 million in 1999. However, compared to Angola, where net inflows of FDIincreased during the same period from US$170 million to over US$2,470 million,Uganda’s increase looks pale (see Figure 2).
Bolivia also shows a steady increase in net inflows in FDI. However, newdisbursements of private non-guaranteed loans by commercial banks to Boliviadecreased after reaching its first decision point under the HIPC Initiative (see Figure 3).Bolivia was also unable to receive any private non-guaranteed bonds.
The trends of private capital flows to Burkina Faso and Guyana are even moredisappointing. Not only did neither country receive any disbursement of private non-guaranteed loans11 or bonds, their net inflows of FDI decreased after reaching their firstdecision points under the HIPC Initiative (see Figure 4).
Figure 2Net inflows in FDI
(current US$ million)
0
500
1,000
1,500
2,000
2,500
1994 1995 1996 1997 1998 1999
Uganda Bolivia Angola
11 With exception that Guyana received a marginal US$27 million from a semi-commercial bank in1999.
7
Figure 3Private non-guaranteed disbursements of commercial banks
(in current US$ million)
269
0 0 0 0 0 2.9
25
6791
127131
0
100
200
300
1994 1995 1996 1997 1998 1999
Bolivia
Sum of Uganda, Burkina Faso, and Guyana
Figure 4Net inflows in FDI
(current US$ million)
107
101013
1710
18
7493
484753
0
20
40
60
80
100
120
1994 1995 1996 1997 1998 1999
Burkina Faso Guyana
2.1.2 Changes in official capital flows
As Figure 5 shows, disbursements of public and publicly guaranteed (PPG) debt to thegroup of HIPCs have been decreasing since 1995, even though they increased sharplyfor the group of non-HIPC low-income countries from 1996-98. Note that the decreaseto the group of HIPCs would be even worse if the data would be expressed in real terms.It is also interesting to look at the longer-term developments among these two groups:note that until 1986, disbursements to HIPCs have been higher than disbursements tolow-income countries excluding HIPCs, but consistently lower since 1987.
8
We next look at the same flows after excluding disbursements from the IMF, wherebywe concentrate on the period of 1994-99, reflecting three years (1994-96) before andthree years (1997-99) after the adoption of the HIPC Initiative. Figure 6 shows that afterexcluding disbursements from the IMF, the decline in disbursements of PPG debt to thegroup of HIPCs happens two years later. In other words, the IMF has started earlier thanother official creditors to decrease its disbursements to the group of HIPCs. However,there still remains a decrease in disbursements to the group of HIPCs from 1997 to 1998and again from 1998 to 1999, even though disbursements increased sharply for thegroup on non-HIPC low-income countries from 1997-98 and decreased less from 1998-99. In real terms (using the SDR interest rate as discount rate), the 1997-99 averagedisbursements of official creditors excluding the IMF to the group of HIPCs have been18 per cent lower than the 1994-96 average disbursements.
Figure 5Disbursements on public and publicly guaranteed (PPG) debt by official creditors, including IMF
(in current US$ million)
0
5
10
15
20
25
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998
HIPCs Low income, excl. HIPCs
Figure 6Disbursements on public and publicly guaranteed (PPG) debt by official creditors, excluding IMF
(in current US$ million)
0
5
10
15
1994 1995 1996 1997 1998 1999
HIPCs Low income, excl. HIPCs
9
Figure 7Bilateral and multilateral disbursements to the group of HIPCs
We can further differentiate official disbursements between bilateral and multilateralcreditors, which are shown in Figure 7. It shows that within the last six years,disbursements from multilateral creditors reached a maximum in 1997, followed by arelatively sharp decrease of more than 22 per cent in nominal terms in 1998. Fordisbursements from bilateral creditors, the maximum of the last six years was reached in1995, followed by two reductions of about 20 per cent per year, a near recovery ofbilateral disbursements to 1995 levels in 1998, and a subsequent reduction in 1999.However, please note that the only reason for the near recovery of bilateraldisbursements in 1998 was due to a sharp increase in bilateral disbursements toVietnam, which increased—related to the Asian Crisis—from US$216 million in 1997to over US$775 million in 1998. Excluding Vietnam, there is a steady decline inbilateral disbursements to the group of HIPCs since 1995.
What makes these overall negative trends worse is that even concessionaldisbursements, especially bilateral concessional disbursements, have decreased and aswill be shown below, that this picture remains valid even after including grants. AsTable 2 shows, the percentage reduction in bilateral concessional disbursements fromthe 1994-96 average to the 1997-99 average was more than 15 per cent in nominalterms.
Given that HIPC debt relief is calculated based on each creditor’s share in NPV debt,these reductions in disbursements to HIPCs are quite rational from each creditor’s pointof view. However, from an overall welfare perspective, these reductions in debt flows toHIPCs are not rational, as they are likely to reduce the growth potentials of HIPCs andthus undermine their long-term debt sustainability. Hence, the situation is a typical‘prisoner’s dilemma’ where rational behaviour of each creditor leads to a sub-optimalmacro outcome.
The next question we want to analyse is what the trends of public and publiclyguaranteed disbursements were for the four HIPCs, which have reached their originaldecision points in 1997. We first compare the bilateral disbursements during the years
10
directly before the adoption of the HIPC Initiative with the bilateral disbursements ofthe early 1990s. Given that under the original framework, each bilateral creditor’s sharein HIPC debt relief was fixed according to its share of NPV debt at the decision pointDSA, the hypothesis is that bilateral disbursements to 1997 decision point HIPCsdecreased shortly before the decision point in order to minimize the costs of the HIPCInitiative. Indeed, as Table 3 shows, the disbursements to Uganda, Bolivia, BurkinaFaso and Guyana have all been reduced sharply (and beyond the reduction to the groupof HIPCs) in the three years before their decision points, even though bilateraldisbursement to other (non-HIPC) low-income countries remained nearly the same.
Table 2Average disbursements of concessional debt to the group of HIPCs
Second, we compare the average bilateral disbursements during the years before andafter the adoption of the HIPC Initiative (1994-96 vs. 1997-99). The two alternativehypotheses are the following:
a) As long as bilateral creditors believe that the HIPC Initiative provides debtsustainability, they have no reason to continue shifting disbursements fromdecision point HIPCs to other countries that are not covered under the HIPCInitiative.
b) However, if bilateral creditors believed that the original HIPC Initiative did notprovide an exit from future debt rescheduling, they would continue to shiftdisbursements from HIPC decision point countries to non-HIPCs, even afterreaching the decision point.
Though there may be other explanations, the data provided in Table 4 seem to generallysupport the hypothesis that bilateral creditors did not believe that the original HIPCInitiative would provide a lasting exit from future debt rescheduling. Furthermore, giventhat the original HIPC framework allocated the multilateral costs of HIPC debt relief
11
according to multilateral shares at the completion point DSA, we can comparemultilateral disbursements before and after the completion point DSAs. Given that (a)both Uganda and Bolivia reached their completion point in 1998, with the completionpoint DSA based on 1997 debt data, and (b) multilateral disbursements to low incomecountries other than HIPCs increased during 1998-99, compared to the 1994-97average, we should also have expected an increase in multilateral disbursements toUganda and Bolivia. However, as Table 5 shows, this was not the case. Again, whilethere may be other explanations, the data support the hypothesis that multilateralcreditors did not believe that Uganda and Bolivia would have achieved debtsustainability after reaching the completion point under the original HIPC framework.12
One possible explanation for these declining trends in official disbursements to HIPCs,and especially to the early decision point HIPCs, could be that donors shifted from loansto grants. We thus look shortly also at changes in grants provided to HIPCs. The usualdistinction made is between grants excluding technical cooperation and technicalcooperation grants. To make a long story short, both categories of grants provided to thegroup of HIPCs during 1997-99 decreased compared with the three years before theadoption of the HIPC Initiative (1994-96): grants excluding technical cooperationdecreased by more than 16 per cent (from an average US$9.4 billion during 1994-96 toan average US$7.8 billion during 1997-99); technical cooperation grants decreased bymore than 18 per cent (from an average US$3.8 billion during 1994-96 to an averageUS$3.1 billion during 1997-99). Looking at the four HIPCs which reached their firstdecision points in 1997, the annual average of total grants before the decision point(1994-97) also decreased compared with the annual average of total grants afterreaching the decision point: Uganda’s average decreased by 0.2 per cent, Bolivia’sdecreased by more than 22 per cent, Burkina Faso’s decreased by 8.8 per cent, andGuyana’s decreased by 9.5 per cent. Note again that all these flows are in nominalterms, thus, in real terms, the reductions are much larger.
Table 4Bilateral disbursements, 1994-99
(current US$ million, if not noted otherwise)
Average Percentage change:
1994-96 1997-99 1994-96 to 1997-99 1990-93 to 1997-99
Note: (a Guyana’s tripling of disbursements has to be seen in relationship to the earlier cuts indisbursements; compared to the early 1990s, disbursements to Guyana are still low.
Source: World Bank (2001a).
12 While it is too early to analyse the implications of the enhanced HIPC initiative, the decisions takenon the delivery of enhanced HIPC debt relief seem to indicate that bilateral and multilateral creditorsare still unconvinced that the enhanced HIPC initiative provides a lasting exit from future debtrescheduling.
12
Table 5Disbursements to Uganda and Bolivia
before and after reaching their completion points (a
Average (current US$ million)
1994-97 1998-99
Percentage change%
Disbursements of multilateral creditors, excluding IMF
Uganda (04/97; 04/98)) 230.8 169.4 -27
Bolivia (09/97; 09/98) 329.7 245.8 -25
HIPCs 4,918 4,109 -16
Disbursements of official creditors, excluding IMF
Uganda (04/97; 04/98)) 254.7 174.6 -31
Bolivia (09/97; 09/98) 375.4 291.6 -22
HIPCs 6,287 5,504 -12
Disbursements of official creditors, including IMF
Uganda (04/97; 04/98)) 312.5 217.2 -31
Bolivia (09/97; 09/98) 410.7 325.9 -21
HIPCs 7,170 6,406 -11
Note: a Though both countries reached their completion point under the original framework of theHIPC Initiative in 1998, the completion point DSAs are based on is the 1997 debt data.
Source: World Bank (2001a).
2.1.3 Changes in private capital flows that are publicly guaranteed
For reasons of completeness, we finally look shortly at disbursements of private debtthat are publicly guaranteed. Within the group of HIPCs, private debt that is publiclyguaranteed is highly concentrated among a few countries: Angola, Ghana, Kenya, andVietnam. In 1999, more than 68 per cent of the private publicly guaranteed debtdisbursed to the group of HIPCs went to Angola (US$875 million), and more than28 per cent went to Ghana, Kenya, and Vietnam (US$359 million). Excluding Angola,the 1998-99 average of disbursements to the group of HIPCs was nearly half of theaverage disbursements during 1994-97. For the same periods, average disbursements toBolivia decreased by 20 per cent (from US$6.1 million to US$4.9 million), newdisbursement to Guyana ceased (compared to an average US$2.4 million during 1994-97), and neither Uganda nor Burkina Faso received any disbursement of private debt(either public guaranteed or non-guaranteed) during 1994-99.
2.1.4 Conclusion on changes in capital flows to HIPCs
Combining the various results so far, there is some indication that (i) the adoption of theHIPC Initiative led to a reduction in disbursements to HIPCs, (ii) HIPC debt relief hasbeen deducted from traditional development assistance, (iii) foreign direct investmentflows to HIPCs are largely concentrated to a few countries and unrelated to a country’sdebt sustainability, and (iv) the already marginal private capital flows to HIPCsgenerally continue to be decreasing. Given that overall development budgets aregenerally decreasing in real terms and that private creditors and investors also faceconstraints from a slowing world economy, it is to expect that this negative tendency ofchanges in capital flows to HIPCs might continue. In other words, HIPCs could end upwith no additional resources for poverty reduction.
13
2.2 Changes in investment and savings ratios
In this sub-section, we first review the implications of the long-term changes ininvestment and savings ratios on the growth sustainability of the four HIPCs with thehighest growth rates of real GDP. We then look at the implications of the most recentchanges in investment and savings ratios on the growth prospects of the four HIPCs thatare expected to grow most during 2000-10. The four HIPCs that grew most during thelast decade were Uganda (6.7 per cent), Mozambique (6.3 per cent), Guyana (6.0 percent) and Benin (4.3 per cent).13 The changes in investment and savings ratios of thesefour countries over the last decade are shown in Figures 8 and 9, respectively.
The DSAs’ average annual growth assumptions of real GDP for these four countriesduring 2000-10 are 5.6 per cent for Uganda, 5.9 per cent for Mozambique, 4.2 per centfor Guyana, and 5.5 per cent for Benin. Note that the growth assumptions for Uganda,Mozambique and Guyana are lower than what these three HIPCs achieved in the lastdecade, but higher for Benin (4.3 per cent versus 5.5 per cent). Looking at Figures 8 and9, there are some indications for the lower growth projections for Uganda and especiallyfor Guyana. On the other hand, neither the recent trends nor the levels of Benin’sinvestment and savings rates justify its higher growth assumptions (which is even higherthan that of Guyana and nearly equal to that of Uganda and Mozambique).
Figure 8Gross domestic investment, 1990-99
10
15
20
25
30
35
40
45
1990 1992 1994 1996 1998
Pec
enta
gesh
are
inG
DP
Uganda (6.7%; 5.6%) Mozambique (6.3%; 5.9%)
Guyana (6.0%; 4.2%) Benin (4.3%; 5.5%)
13 Mauritania also grew at an average 4.3 per cent during the 1990s, but is analysed in the group of fourHIPCs, which are supposed to grow most during 2000-10.
14
Figure 9Gross domestic savings, 1990-99
-25
-15
-5
5
15
25
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Uganda (6.7%; 5.6%) Mozambique (6.3%; 5.9%)
Guyana (6.0%; 4.2%) Benin (4.3%; 5.5%)
Figure 10Gross domestic investment, 1990-99
5
10
15
20
25
30
35
40
45
50
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Mauritania (4.3%, 7.3%)
Guinea-Bissau (0.3%, 7.0%)
Madagascar (1.8%, 6.2%)
Rwanda (-1.6%, 6.1%)
15
Figure 11Gross domestic savings, 1990-99
-25
-15
-5
5
15
25
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Mauritania (4.3%, 7.3%)
Guinea-Bissau (0.3%,7.0%)
Madagascar (1.8%, 6.2%)
Rwanda (-1.6%, 6.1%)
The picture is even worse for the four countries which are expected to grow most during2000-10: Mauritania (7.3 per cent), Guinea-Bissau (7.0 per cent), Madagascar (6.2 percent) and Rwanda (6.1 per cent), whereby we concentrate on the most recent trends(1997-99) of the investment and savings ratios to check if there is some indication forthe high growth assumption of more than 6 per cent (see Figures 10 and 11).Mauritania’s investment ratios are stagnating and its savings ratios have declined duringthe last three years (1997-99). Guinea-Bissau’s investment and savings ratios werelower in 1999 than in 1997, Madagascar’s savings and investment ratios have onlymarginally improved during the last three years (1997-99), Rwanda’s investment ratiohas been lower in 1999 than it was in 1997 and its savings ratios have remainednegative since 1993. Please see Appendix Table for the graphical illustrations of theinvestment and savings ratios as well as the structural transformations of all 22 HIPCsthat reached the enhanced decision point by end-December 2000, with furtherinformation on the data sources.
2.3 Structural transformation
A similar picture emerges from looking at the structural transformation of the four highgrowth HIPCs of the 1990s (see Figure 12). While Uganda and especially Mozambiqueshow a positive structural transformation, Guyana’s and Benin’s structural trans-formations have stagnated for the last seven years. Furthermore, given that there is usuallyno advantage of structural backwardness, looking at the levels of manufacturing shares inGDP, it seems unlikely that Benin will be able to achieve the assumed growth rates.
Finally, looking at the relatively stagnant structural transformation during the last threeyears of Mauritania, Guinea-Bissau, Madagascar, and Rwanda (see Figure 13), there isno structuralist foundation for the highly optimistic growth rates of more than 6 per centfor these four countries.
16
Figure 12Structural transformation, 1990-99
4
8
12
16
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Uganda (6.7%; 5.6%) Mozambique (6.3%; 5.9%)
Guyana (6.0%; 4.2%) Benin (4.3%; 5.5%)
Figure 13Structural transformation, 1990-99
4
8
12
16
1990 1992 1994 1996 1998
D Mauritania (4.3%; 7.3%) Guinea-Bissau (0.3%; 7.0%
Madagascar (1.8%; 6.2%) Rwanda (-1.6%; 6.1%)
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
2.4 AIDS, climate changes, and armed conflicts
The IMF and World Bank (2001: 7) study mentions that ‘longer-term growth prospectscan be undermined by natural disasters, war or health threats such as the AIDS epidemicaffecting many of the HIPCs’. However, the report fails to analyse the impacts of thesefactors on growth. It also fails to recognize that these factors are not only possibilitiesbut also realities. A factor that is not reality yet, but which Lindert and Williamson(2001) have called a ticking bomb is possible south-north mass migration, especially out
17
of Africa,14 which would also have a negative impact on the nominal growth rate ofsouthern exports. A further possibility is that the terms of trade will continue to show along-term negative trend, while most DSA assumptions are that HIPCs’ terms of tradewill recover.
With regards to AIDS, current estimates of economic impacts translate to annualreductions in real GDP growth rates between 0 and 3 percentage points, depending on acountry’s HIV prevalence (ranging up to 35.8 per cent at end-1999 in the case ofBotswana).15 Note that losses in annual per capita growth are much lower since manyof the infected people are assumed to die. In the words of a recent World Bank (2000: 9)AIDS study:
The illness and impending death of up to 25 per cent of all adults in somecountries will have an enormous impact on national productivity andearnings. Labour productivity is likely to drop, the benefits of educationwill be lost, and resources that would have been used for investmentswill be used for health care, orphan care, and funerals. Savings rates willdecline, and the loss of human capital will affect production and thequality of life for years to come.
Obviously, all these impacts will also affect the HIPCs’ export growth rates, which arethe key debt sustainability indicator within the HIPC framework.
With regards to natural disasters, the number of hydro-meteorological disasters hasmore than doubled over the second half of the last decade. In 2000, the InternationalFederation of Red Cross and Red Crescent Societies (IFRC) registered 752 naturaldisasters, versus 609 in 1999 and 481 in 1998.16 Predictions are that this trend is likelyto accelerate,17 whereby agricultural production in many tropical and subtropicalcountries, especially in Sub-Saharan Africa and Latin America (hence mostly in HIPCs)is likely to decrease. Furthermore, even less dramatic weather changes can havesignificant impacts on export growth. For example, Uganda’s reassessment under theenhanced HIPC Initiative has shown that the projections on NPV debt and exports madein Uganda’s original completion point document were too optimistic.18
14 See Hatton and Williamson (2001).
15 See IMF Survey of November 6, 2000, which reports that for Swaziland (which had an HIVprevalence rate of 25 per cent at end-1999) the IMF staff estimates that, by 2010, the rate of GDPgrowth will be about 2 per cent lower than it would be if Swaziland were not experiencing an AIDSepidemic.
16 For the IFRC, a disaster means that at least 10 people died, 100 people were affected, and a state ofemergency was declared or international aid sought; see Reuters Business Briefings, June 28, 2001.
17 See the United Nations’ third assessment on climate change of July 2001 which states that globaltemperatures are rising nearly twice as fast as previously thought; see also the recent statements byRobert Watson, World Bank Chief Scientist and Chairman of the Intergovernmental Panel of ClimateChange.
18 As reported in Uganda’s enhanced decision point document of January 2000 (available on the HIPCwebsite: www.worldbank.org/hipc) lower commodity prices and adverse weather conditions causedan unexpected decrease of exports of nearly 25 per cent in fiscal year 1998.
18
With regards to war and civil conflicts, it is unlikely to assume that all these internal andexternal conflicts will cease after having reached the HIPC decision or completionpoints.19 Given that these conflicts have generally a devastating impact on growth,especially if prolonged, realistic growth projections need to account for the economicimpact of such conflicts. For example, though it was already known in December 2000that Guinea’s growth rate for 2000 would be less than two per cent due to the armedconflict at its southern borders, the December 2000 DSA for Guinea neverthelessassumed a growth rate of more than 4 per cent for year 2000. Finally, the delay in Côted’Ivoire’s enhanced decision point shows how real conflicts are even in cases where acountry has reached a HIPC decision point.
3 HIPC debt sustainability indicators
As Ajayi and Khan (2000) have pointed out recently, given that it is in practice nearlyimpossible to calculate the sustainable level of foreign borrowing, various ratios—suchas that of debt to exports, debt service to exports, and debt to GDP (or GNP)—havebecome standard measures of sustainability. However, the HIPC framework definesdebt sustainability largely by a debt-to-export ratio, whereby under the enhancedframework, a NPV debt-to-export ratio of 150 per cent (down from 200-250 per cent) isconsidered to be sustainable. Only for countries having an export-to-GDP ratio of atleast 30 per cent (down from 40 per cent) and a government revenue-to-GDP ratio of atleast 15 per cent (down from 20 per cent), a NPV debt-to government revenue ratio of250 per cent (down from 280 per cent) is considered to be sustainable. Of the23 countries having reached the decision point under the enhanced HIPC Initiative byend-June 2001, only four countries (Guyana, Honduras, Mauritania, and Senegal) havequalified under the fiscal criteria. This section analyses first the appropriateness of theHIPC debt sustainability indicators based on the recent literature and provides thensome empirical evidence on the relevance of various debt sustainability indicators.
3.1 Appropriateness of the HIPC debt sustainability indicators
One of the most serious critiques of the HIPC framework is that it uses inappropriate orat least insufficient debt sustainability criteria. For example, Sachs (2000) has expressedthe view that the HIPC sustainability criteria have nothing to do with debt sustainabilityin any real sense. Others have stressed that ‘the ratios of debt and debt service toexports, which are more frequently used, are hard to justify on theoretical grounds’, andthat ‘at the very least, indicators relative to GDP should be taken as seriously asindicators relative to exports’.20 Finally, there is some doubt if the NPV calculationsused in the HIPC framework are appropriate. Among many problems related todiscounting, the key argument is that discounting unpayable debt at market discountrates gives the wrong picture about a HIPC’s debt burden.21
19 For a theoretical and empirical analysis of the effects of economic policy and the receipt of foreign aidon the risk of civil war, see Collier and Hoeffler (2000).
20 See IMF (1998: 39-40).
21 See Gunter (forthcoming) for some further details.
19
While the debt-to-export ratio has some justification for the determination of an upperlimit of a country’s debt sustainability, it says very little about a government’s ability torepay its external public debt. As a World Bank (2001b: Table 3) HIPC study shows, atleast four countries (Guinea, Mauritania, Niger, and Zambia) will continue to paybetween 20-23 per cent of their government revenues as external debt service on publicor publicly guaranteed debt after enhanced HIPC debt relief. While this is largely due tothe highly restrictive thresholds for the application of the HIPC fiscal indicator, a moreflexible debt-to-export criterion could also avoid such problems.22
In any case, as is well-known, the debt-to-export ratio has been used for mostly middle-income Latin American countries in the aftermath of the 1982 debt crisis, whereby asubstantial part was private debt and exchange rate adjustments ensured substantialtrade surpluses. However, most HIPCs import not only more than they export(Cameroon and Côte d’Ivoire are two exceptions). Indeed, as was shown for example byLópez and Thomas (1990), Sub-Saharan African economies depend highly on imports.Furthermore, due to access constraints to industrialized countries’ markets fordeveloping country exports as well as some market saturations for HIPC exports, tradedeficits are likely to remain for HIPCs at least over the next 10 to 20 years.23
Even if HIPCs would be forced to cut their imports and increase their exports to earn theforeign exchange needed to repay external debt, the government owing the debt mayonly get a small amount of the export revenues.24 For example, multinationalenterprises own close to 90 per cent of Guinea’s exports and use most of the foreignexchange earnings for imports of equipment, salaries of expatriate workers, andtransfers of profits. In some cases, exports of HIPCs reflect a large degree of re-exports(the exports simply pass through the country and no foreign exchange is earned byanybody).25 Finally, the way the NPV debt-to-export criterion is currently used in theHIPC framework discourages a HIPC’s export-led growth strategy, especially in HIPCswhere the decision point is some time in the future.26
Related to the inappropriate debt sustainability indicator is the critique that the HIPCInitiative’s country selection is too narrow.27 Some of the world’s poorest countries
22 Indeed, Cohen’s (1996) analysis has shown that the sustainable debt-to-export ratios associated to a 25per cent discount of estimated secondary market prices would be 68 per cent for Guinea, 90 per centfor Niger, and 79 per cent for Zambia.
23 See Hersel (1998) for a more detailed analysis of sustained trade deficits and the interdependencybetween external debt and trade policy.
24 These budgetary aspects of the transfer problem have been analysed by Dani Rodrik and especially byHelmut Reisen, see Hjertholm (1999) for further references.
25 Thus far, the HIPC framework has not been consistent in either including or excluding re-exports inthe calculation of the debt-to-export criteria.
26 Though some precaution had been taken to capture export volatility by defining exports as a three-year backward looking average, it would have been better to take a much longer backward lookingaverage ending with the year before the framework of the initiative is adopted, not with year previousto the HIPC decision point.
27 The HIPC initiative defines a country as ‘heavily indebted’ if traditional debt relief mechanisms areunlikely to reduce a country’s external debt to a sustainable level. The criterion for being ‘poor’ is tobe an IDA-only country. An IDA-only is considered to rely on financial resources from the WorldBank’s International Development Association (IDA), whereby the main criterion is based on acountry’s GDP per capita. However, a few other exceptions have been made to this income per capita
20
have been excluded from the HIPC Initiative as their debt is—according to the narrowlydefined HIPC criteria—considered to be sustainable. In the words of a recentEURODAD (2001) report, the concept of debt sustainability has to be approached froma human and social development perspective. The fact that IDA-only countries likeBangladesh, Cambodia, Haiti, Nepal, and Tajikistan have a GDP per capita of less thanone-dollar-a-day, makes it obvious that these countries do not have their own resourcesto repay their external debt, not now and not in the foreseeable future.28 The onlyreason these poor countries can service their external debt currently is that they receivecurrently new loans that are more than sufficient to repay old debt. However, thiscontinuous repaying of old debt with new debt cannot be considered to constitute debtsustainability.
There are also cases in which low-income heavily indebted countries are not part of theHIPCs, as they are considered to be not IDA-only countries. The example of Nigeria isrelatively well-known. Nigeria is highly indebted: its NPV debt-to-export ratio isestimated to be 188 per cent (38 per cent higher than what is considered to besustainable under the enhanced framework). It is also a poor country: GDP per capita isbelow US$300 a year and more than 70 per cent of its 120 million population live inabsolute poverty (below one-dollar-a-day). The official reason for Nigeria’s exclusionfrom the group of HIPCs is that Nigeria does—due to its large oil reserves—not rely onIDA assistance. In other words, it is expected that Nigeria extracts and sells its oil togenerate the foreign exchange and revenues necessary to repay its external debt, most ofwhich was contracted by corrupt previous governments.
3.2 Empirical evidence on the relevance of debt sustainability indicators
While there are overwhelming theoretical arguments for extending the debtsustainability indicators of the HIPC Initiative, little empirical evidence for the relativeimportance of various debt sustainability indicators has been presented so far.29 Weintend to draw some conclusions on the relative importance of debt sustainabilityindicators by substituting three debt sustainability indicators in otherwise standardspecifications of a macroeconomic investment function.30 The three debt sustainability
criterion. Originally, Nigeria and Equatorial Guinea were also considered to be HIPCs, but have beendropped from the list of HIPCs as they were later on considered to be no more IDA-only eligiblecountries. Malawi and The Gambia have been added recently as it became clear that their debt ishigher than initially estimated. For the most current list of HIPCs, see the HIPC website:www.worldbank.org/hipc/
28 Even if these countries would have an income per capita growth rate of 5 per cent per year (none ofthem did so in the past or do so currently), they would still remain to be IDA-only countries in 20years.
29 See Hjertholm (1999) for an excellent review of the analytical history of HIPC debt sustainabilitytargets, whereby he points out that there is no analytical basis for the appropriate level of the HIPCfiscal indicator.
30 The traditional investment literature has gone through considerable changes from the acceleratortheory, to the neoclassical theory of investment and finally, Tobin's q-theory. However, based on thepoor empirical performance of these traditional investment theories, recent research of the investmenttheory has led to a revised and extended account of the determinants of investment. This holdsparticularly true for the determinants of investment in developing countries. A good review ofempirical investment function specifications for developing countries is provided by Rama (1993).The specification here broadly follows that of Gunter (1998) and Oshikoya (1994).
21
indicators analysed are (i) the debt-to-export ratio [DEXP], (ii) the debt-to-GNP ratio[DGNP], and (iii) the debt-to-government revenue ratio [DREV]. In algebraic terms, weestimate the following three regressions:31
INVFIPR = the ratio of fixed private domestic investment to GDP,
INTLEN = the nominal lending interest rate,32
HESLAG = the lagged real growth rate (based on PPP adjusted GDP),33
Given that the Durbin-Watson statistics of the initial regressions ranged from 0.31-0.34(indicating autocorrelated errors), an AR(1) specification has been applied to theregressions to take care of first order autocorrelation. After invoking the AR(1) errorcorrection, the Durbin Watson statistics range from 1.7 to 2.2, thus indicating no furtherautocorrelation.
31 All data come from the Branson, Guerrero and Gunter (forthcoming) database, a database compiledmainly from various World Bank and IMF databases but also from other sources, like the Penn WorldTables (the Summers and Heston database). The database covers 25 years of time-series data of 21industrialized and 72 developing countries (including 30 HIPCs), whereby the number of countriesincluded in the databases has been limited to have a near balanced panel. Since data on debt indicatorsare limited to developing countries, we have excluded the 21 industrialized countries, leaving us—dueto other constraints on data availability—with a minimum of at least 1000 observations for eachvariable.
32 The usual proxy for the cost of capital is the real interest rate. However, given that data on real interestrates are highly distorted, many earlier investment studies have shown that real interest rates arehardly a significant determinant of investment. This is consistent with the view that the cost of capitalis determined by other factors besides interest rates, they are not easy to get a hold of. Also, instead ofusing inflation biased real interest rates, nominal interest rates may serve as a better proxy of both thecost of capital and the availability of credit.
33 There is little doubt that the income accelerator is a considerable determinant of investment. Onesimple and good measure of the income accelerator is the current growth rate of GDP. However, giventhe likely bivariate causality between the current growth rate and investment, the growth rate needs tobe lagged by one period. It is obviously real, not nominal growth, which is the more appropriatevariable in the determination of investment. The lagged real growth rate serves also as broadapproximation of the availability of investment funds, which constitute another important determinantof investment.
22
Table 6Regression results
LOG
(INTLEN) HESLAG FDI FINDEV1 FINDEV2
LOG
(DEXP)
LOG
(DGNP)
LOG
(DREV)
Part 1: Simple basic specification
(R-squared varies between 0.80 and 0.82; DW statistic varies between 1.74 and 2.13)
EQ. 1.a -0.26 0.04 -1.84
t-stat -0.4 1.9 -2.4
EQ. 1.b -0.03 0.03 -1.92
t-stat -0.1 1.8 -2.5
EQ. 1.c -0.18 0.05 -2.59
t-stat -0.3 2.4 -3.2
Part 2: Specification with foreign direct investment
(R-squared varies between 0.80 and 0.81; DW statistic varies between 1.74 and 2.17)
EQ. 2.a -0.66 0.04 0.17 -2.24
t-stat -1.1 1.7 1.5 -2.7
EQ. 2.b -0.42 0.04 0.26 -2.80
t-stat -0.7 1.7 2.3 -3.3
EQ. 2.c -0.46 0.06 0.21 -3.02
t-stat -0.7 2.4 1.6 -3.4
Part 3: Specification with financial market development indicator #1
(R-squared varies between 0.81 and 0.82; DW statistic varies between 1.75 and 2.19)
EQ. 3.a -0.47 0.04 0.18 7.40 -1.85
t-stat -0.8 1.8 1.6 1.7 -2.2
EQ. 3.b -0.26 0.04 0.27 7.79 -2.58
t-stat -0.4 1.7 2.4 1.8 -3.0
EQ. 3.c -0.37 0.07 0.21 6.54 -0.17
t-stat -0.5 2.6 1.7 1.3 -3.1
Part 4: Specification with financial market development indicator #2
(R-squared varies between 0.81 and 0.82; DW statistic varies between 1.73 and 2.16)
EQ. 4.a -0.68 0.04 0.17 4.70 -2.02
t-stat -1.1 1.6 1.5 2.2 -2.4
EQ. 4.b -0.48 0.03 0.26 4.75 -2.60
t-stat -0.8 1.5 2.2 2.2 -3.0
EQ. 4.c -0.55 0.06 0.21 5.08 -2.75
t-stat -0.8 2.3 1.6 2.2 -3.1
The results for the three basic regressions (1.a to 1.c) are presented in Part 1 of Table 6.It turns out that—among the three debt sustainability indicators—the debt-to-government revenue ratio (DREV) is the most significant determinant for private fixedinvestment, followed by the debt-to-GNP (DGNP) and debt-to-exports (DEXP) ratios.
We have tested the robustness of the results with a variety of alternative investmentspecifications, all of which provide the same results with regards to the relativeimportance of the three debt sustainability indicators. First, given the importanceforeign direct investment plays as a catalyst for private investment in developing
23
countries, we have added net foreign direct investment (in per cent of GDP) to themodel specification. It has also been suggested that FDI is a better indicator than thepartly complementary, partly substitutionary public fixed investment. Second,recognizing that the recent literature34 has shown that financial market development isone of the most robust determinants for investment and growth, we have added twostandard indicators for financial marked development: (i) the ratio of liquid liabilities ofthe financial system to GDP, and (ii) the ratio of deposit money bank domestic assets todeposit money bank domestic assets plus central bank domestic assets. While the moredetailed results of these alternative model specifications are presented in Parts 2-4 ofTable 6, the important point is that the relative significance of the three debtsustainability indicators remains the same, whatever the change in the modelspecification.35
4 Some possible modifications
While more comprehensive suggestions on necessary improvements and possiblemodifications of the HIPC Initiative are presented in Gunter (forthcoming), the goal ofthis section is to make some more specific suggestions on issues related to the HIPCInitiative’s long-term debt sustainability.
4.1 Aid coordination, additionality, and burden-sharing
First of all, more aid coordination is needed to reverse the currently decreasing trends indevelopment assistance, especially to HIPCs. More specifically, to avoid the currentprisoner’s dilemma for new disbursements to HIPCs that have not reached their decisionpoint, the HIPC Initiative’s burden-sharing should be based on a simplified end-2000DSA. The discount rates used for such an end-2000 DSA, as well as any future DSA,should be low, fixed, and uniform across currencies (e.g. 3 per cent). The burden-sharing concept should also be enhanced, by taking the creditor’s economic power intoaccount.
4.2 Realistic growth assumption
Second, the DSA growth assumptions should be more realistic. This could be achievedby projecting future growth rates based on macroeconomic principles like analysing
34 See Levine (1997) for a review and synthesis.
35 We have also tested the robustness of the relative significance of the three debt sustainabilityindicators by substituting the three debt ratios (DEXP, DGNP, and DREV) with three correspondingdebt service ratios: (i) the debt service-to-export ratio [DSEXP], (ii) the debt service-to-GNP ratio[DSGNP], and (iii) the debt service-to-government revenue ratio [DSREV]. The rational is thatsubstituting the three debt ratios with corresponding debt service ratios should give us the same resultsin terms of relative importance of three indicators. However, we should keep in mind that privateinvestors are more likely to look at the implications of the debt stock on future debt service paymentsthan at the current debt service payments. Thus, we should expect lower significance levels for thethree debt service indicators. Indeed, our regression results show that substituting the debt ratios withcorresponding debt service ratios results in insignificant t-statistics for the three debt service ratios,however, the message on the relative importance of the three indicators remains valid.
24
recent trends in structural transformations, investment and savings rates, as well as byaccounting explicitly for negative impacts related to AIDS and climate change.
4.3 Gradual replacement of loans by grants?
Third, following a suggestion made by the US International Financial InstitutionAdvisory Commission, commonly referred to as the Meltzer Commission,36 discussionis currently underway to gradually replace new IDA loans to the poorest countries bygrants. Though this would obviously be favourable for the long-term debt sustainabilityof these countries, the biggest problem is that this could reduce the overall availabilityof financial assistance to the poorest, especially if annual IDA contributions are notincreased appropriately.37 In any case, a selected, gradual and fully-funded replacementshould be analysed in more detail.
4.4 Debt sustainability indicators and HIPC eligibility
Finally, it is suggested to eliminate the two threshold ratios for the applicability of thefiscal debt sustainability indicator (i.e., the requirements of having an export-to-GDPratio of at least 30 per cent and a government revenue-to-GDP ratio of at least 15 percent). While more analysis would be needed to determine the appropriate level of thefiscal indicator, whereby a few country-specific vulnerability factors (e.g., exportconcentration) should be taken into account, the recommendation to focus less onexport-related and more on government revenue-related indicators is not new. Forexample, more than ten years ago, Dittus (1989) had analysed the budgetary dimensionof the debt crisis in low-income Sub-Saharan Africa and suggested to assign the debtservice-to-revenue ratio a central role.
Furthermore, it is suggested to replace the current ‘IDA-only’ requirement with apurchasing-power parity (PPP) based GDP per capita level, whereby two categories aresuggested: category-one countries would have a PPP-based GDP per capita belowUS$1,500;38 category-two countries would have a PPP-based GDP per capita betweenUS$1,500 and US$3,000.39 For category 1 countries it is suggested to limit debt servicepayments to 5 per cent of the average 1990-99 government revenues, independent ofwhat the NPV debt-to-export and the NPV debt-to-government revenue ratios are. The
36 A slightly edited version of the report is available at this University of Michigan site:www.econ.lsa.umich.edu/~alandear/topics/meltzer.html .
37 On the other hand, there would be no need to worry about such an outcome if all OECD countrieswould raise their aid target set at 0.7 per cent of GNP by the UN many years ago, or if theinternational community could agree to suggestion related to the international taxation of exchangerate speculations. For other problems related to the replacement suggestion, see the US Treasury’sresponse to the Meltzer report, available at: http://www.ustreas.gov/press/releases/docs/response.pdf .
38 Category one would include the 31 poorest HIPCs, but also add Bangladesh, Bhutan, Cambodia,Comoros, Djibouti, Eritrea, Haiti, Nepal, Nigeria, and Tajikistan.
39 Category two would include eight of the 10 richest HIPCs (in terms of PPP-based GDP per capita:Bolivia, Cameroon, Côte d'Ivoire, The Gambia, Ghana, Guinea, Honduras, Mauritania, Nicaragua,and Vietnam) and 15 non-HIPCs with a PPP-based GDP per capita between US$1,500 and US$3,000.The only two HIPCs excluded in both categories are Angola and Guyana, which have a PPP-basedGDP per capita of US$3,200 and US$3,600, respectively.
25
decision on which creditor would get paid would need to be based on a predeterminedrule that combines (i) the share of debt service due to each creditor with (ii) thecreditor’s ability to provide debt relief based on the creditor’s income level.
For the richer (though still poor) category-two countries, the current eligibility criteriacould be applied after elimination of the threshold requirements for the fiscal indicator.If eligible, debt service payments should be limited to 10 per cent of the average1990-99 government revenues. These limits on the annual debt service to governmentrevenue ratios would not only ensure that governments could spend their revenues ondevelopment tasks, they would also ensure private investors that the government willnot be forced to increase current and future taxes to effect the repayment of externaldebts. Hence, it would eliminate a very critical determinant of the debt overhang effect.
5 Summary and conclusion
One of the most serious problems of the HIPC Initiative is that it may not achieve itskey goal of providing a solid exit from future debt rescheduling. In trying to assess howlikely HIPCs are to achieve debt sustainability, we first looked at various capital flowsto HIPCs. Excluding a very few exceptions, it was shown that private capital flows aswell as disbursements of private and public creditors to HIPCs have generally decreasedsince the adoption of the HIPC Initiative. Three aspects make these negative trends incapital flows to HIPCs worse: (i) capital flows to non-HIPC low income countries haveincreased (at least in nominal terms), (ii) capital flows continued to decrease forUganda, Bolivia, Burkina Faso, and Guyana, even after they reached their originaldecision points, and (iii) there are even reductions in grants and disbursements ofconcessional loans. These negative trends in terms of capital flows to HIPCs are alsoreflected in an overall reduction in net aggregate resource flows to HIPCs (see Figure14), which imply that compared to previous years, HIPCs had less and less resources fordevelopment tasks at least until 1999. Projections are that net aggregate resource flowsto HIPCs may—due to HIPC debt relief—increase in 2001 and may remain relativelystable after 2001.
Second, we have looked at changes in investment and savings rates and sectoraltransformations of HIPCs during the last decade, which generally indicated that there isno macroeconomic foundation for the high DSA growth assumptions of Guinea-Bissau,Madagascar, Mauritania and Rwanda. A similar analysis for the other HIPCs (providedin Appendix 1) shows that most of the other growth assumptions (excluding Guyana,Mozambique, and Uganda) are also highly optimistic. We have also noted that theimpacts of AIDS and climatic changes have been neglected in the growth assumptionsof most HIPCs.
Third, based on the theoretical and empirical examination of the HIPC debtsustainability criteria, there is some justification to the critique that the HIPC Initiativeuses inappropriate and insufficient debt sustainability indicators. One shoe does not fitall and the thresholds for the fiscal indicator should be eliminated. We have also madesuggestions with regards to limiting HIPC debt service in terms of debt service-to-government ratios under a second enhancement of the HIPC Initiative that differentiatesaccording to PPP-based income per capita levels.
26
Figure 14Net aggregate resource flows to HIPCs
(current US$ million)
17,000
17,500
18,000
18,500
1994 1995 1996 1997 1998 1999
Based on this background the overall conclusion is that the HIPC Initiative is unlikely toprovide a solid exit from future debt rescheduling for many of the poorest countries.Indeed, even the IMF and World Bank acknowledge that the NPV debt-to-export ratio isprojected to remain above 150 per cent for 10 years or more for at least three HIPCs(Bolivia, Malawi, and Niger).40 Note that even there, optimistic export growthprojections have been used for Bolivia (9.1 per cent per annum for 2000-10, comparedwith 5.7 per cent per annum for 1990-99) and Niger (5.4 per cent per annum for 2000-10, compared with –3.9 per cent per annum for 1990-99).41 As long as debtsustainability of many of the poorest countries is questionable, some debt overhangeffects are likely to remain and have a negative impact on investment, growth andpoverty reduction. Though the enhanced HIPC Initiative is likely to have a positiveimpact on poverty reduction in many HIPCs, more sustainable development and furthergoing poverty reduction could be achieved with a more definitive exit fromunsustainable debt than the enhanced HIPC Initiative provides.
40 IMF and World Bank (2001: 19).
41 IMF and World Bank (2001: 24, Table 5).
27
Appendix table
Trends in structural transformation, gross domestic investment rates, and gross domestic savings ratesfor the 22 HIPCs that reached the enhanced decision point by end-December 2000
Benin: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Bolivia: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Burkina: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Cameroon: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Gambia: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Guinea: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Benin: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
PBolivia: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Burkina: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Benin: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Bolivia: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Burkina: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
PCameroon: Gross domestic investment, 1990-99
515253545
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Cameroon: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Gambia: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Gambia: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Guinea: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Guinea: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Appendix table continues….
28
Appendix table (con’t)Trends in structural transformation, gross domestic investment rates, and gross domestic savings rates
for the 22 HIPCs that reached the enhanced decision point by end-December 2000
Guinea-Bissau: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Guyana: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Honduras: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Madagascar: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Malawi: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Mali: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Guinea-Bissau: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Guinea-Bissau: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Guyana: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Guyana: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Honduras: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Honduras: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
PMadagascar: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Madagascar: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Malawi: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Malawi: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Mali: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Mali: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Appendix table continues….
29
Appendix table (con’t)Trends in structural transformation, gross domestic investment rates, and gross domestic savings rates
for the 22 HIPCs that reached the enhanced decision point by end-December 2000
Mauritania: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Mozambique: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Nicaragua: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Niger: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Rwanda: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Mauritania: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Mauritania: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Mozambique: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Mozambique: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Nicaragua: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Nicaragua: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Niger: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Niger: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Rwanda: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Rwanda: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Appendix table continues….
30
Appendix table (con’t)Trends in structural transformation, gross domestic investment rates, and gross domestic savings rates
for the 22 HIPCs that reached the enhanced decision point by end-December 2000
Source: World Bank, World Development Indicators, except forGuinea-Bissau's and Madagascar's 1998 and 1999 percentage shares ofmanufacturing, which have been updated from more recent IMF country reports (available on the IMF website).
Sao T&P: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Senegal: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Tanzania: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Uganda: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Zambia: Structural transformation, 1990-99
0
5
10
15
20
25
1990 1992 1994 1996 1998
Per
cent
age
shar
eof
man
ufac
turin
gin
GD
P
Sao T&P.: Gross domestic investment, 1990-99
5152535455565
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Sao T&P.: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Senegal: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Senegal: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Tanzania: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Tanzania: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Uganda: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Zambia: Gross domestic investment, 1990-99
5
15
25
35
45
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Uganda: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Zambia: Gross domestic savings, 1990-99
-50
-30
-10
10
30
1990 1992 1994 1996 1998
Per
cent
age
shar
ein
GD
P
Appendix-Summary Analysis
Given that we are interested to draw conclusions on the macroeconomic foundation of the DSA growth assumptions,we define a positive trend in the three categories as having realized (1) a total increase between 1990 and 99 of atleast 1 per cent, and (2) a total increase of at least 0.2 per cent during the last 2 years (1997-99). Based on theseminimal criteria for a positive trend, of the 22 HIPCs that have reached the enhanced decision point by end-December 2000:
− only 5 countries experienced a positive trend in their structural transformation (Burkina Faso, Guyana, Mozambique,Rwanda, and Uganda);
− only 7 countries experienced a positive trend in their share of investment to GDP (Burkina Faso, Cameroon, Honduras,Mozambique, Nicaragua, Senegal, and Uganda; and
− only 5 countries have experienced a positive trend in their share of investment to GDP (Benin, Mozambique, Niger, SaoTome & Principe, and Senegal.
Finally, note also that within this group of 22 HIPCs, only Mozambique experienced a positive trend in all threecategories.
31
References
Ajayi, S. Ibi, and Mohsin S. Khan (2000).External Debt and Capital Flight in Sub-Saharan Africa. Washington, DC: International Monetary Fund.
Berthélemy, Jean-Claude, and Ludvig Söderling (2001). ‘The Role of CapitalAccumulation, Adjustment and Structural Change for Economic Take-Off: EmpiricalEvidence from African Growth Episodes’.World Development, 29 (2): 323-43.
Branson, William H., Isabel M. Guerrero, and Bernhard G. Gunter (forthcoming).Macroeconomics for Developing Countries.
Bretton Woods Committee (2000). ‘Concluding Report of a Roundtable Discussion onReassessing Debt Relief, November 2000’. Available at: www.brettonwoods.org .
Cohen, Daniel (1996). ‘The Sustainability of African Debt’. WB Policy ResearchWorking Paper No. 1621 (July). Washington, DC: World Bank.
Collier, Paul, and Anke Hoeffler (2000). ‘Aid, Policy and Peace’. Mimeo. Available at:www.worldbank.org/research/conflict/papers.htm .
Dittus, Peter (1989). ‘The Budgetary Dimension of the Debt Crisis in Low-Income Sub-Saharan Countries’.Journal of Institutional and Theoretical Economics, 145 (2):358-66.
Easterly, William, and Ross Levine (2000). ‘It’s Not Factor Accumulation: StylizedFacts and Growth Models’. WB Policy Research Working Paper. Washington, DC:World Bank.
EURODAD (2001). ‘Debt Reduction for Poverty Eradication in the Least DevelopedCountries: Analysis and Recommendation on LDC Debt’. Brussels: EURODAD.Available at: www.eurodad.org .
Gunter, Bernhard G. (1998). ‘Economic Structure and Investment under Uncertainty’.Washington, DC: American University, Department of Economics. Unpublisheddissertation.
Gunter, Bernhard G. (forthcoming). ‘What’s Wrong With the HIPC Initiative andWhat’s Next?’Policy Development Review. Spring 2002.
Hatton, Timothy J., and Jeffrey G. Williamson (2001). ‘Demographic and EconomicPressure on Emigration out of Africa’. NBER Working Paper No. 8124. Cambridge,MA: National Bureau of Economic Research.
Hersel, Philip (1998). ‘The London Agreement of 1953 on German External Debt:Lessons for the HIPC-Initiative’, in EURODAD,Taking Stock of Debt CreditorPolicy in the Face of Debtor Poverty. Brussels: EURODAD.
Hjertholm, Peter (1999). ‘Analytical History of Heavily Indebted Poor Country (HIPC)Debt Sustainability Targets’. Copenhagen: Development Economics Research Group(DERG), Institute of Economics, University of Copenhagen. Available atwww.econ.ku.dk/derg .
International Monetary Fund (IMF) (1998). ‘Report of the Group of IndependentPersons Appointed to Conduct an Evaluation of Certain Aspects of the Enhanced
32
Structural Adjustment Facility’. Washington, DC: IMF. Available at:www.imf.org/external/pubs/ft/extev/index.htm .
International Monetary Fund (IMF) and World Bank (2001) ‘The Challenge ofMaintaining Long-Term Debt Sustainability’. Progress Report presented to theDevelopment Committee and the International Monetary and Financial Committee(IMFC), Washington, DC: International Monetary Fund and World Bank. Availableat: www.worldbank.org/hipc .
Krugman, Paul (1988). ‘Financing vs. Forgiving a Debt Overhang’.Journal ofDevelopment Economics, 29 (3): 253-68.
Krugman, Paul (1994). ‘The Myth of Asia’s Miracle’.Foreign Affairs, 73 (6) 62-78.
Lindert, Peter H., and Jeffrey G. Williamson (2001). ‘Globalization and Inequality: ALong History’. Paper presented at the Annual Bank Conference on DevelopmentEconomics-Europe, June. Available at: www.worldbank.org/abcde-europe .
López, Ramón E., and Vinod Thomas (1990). ‘Import Dependency and StructuralAdjustment in Sub-Saharan Africa’.World Bank Economic Review, 4 (2): 195-207.
Oshikoya, Temitope W. (1994). ‘Macroeconomic Determinants of Domestic PrivateInvestment in Africa: An Empirical Analysis’.Economic Development and CulturalChange, 42 (3): 573-96.
Rama, Martin (1993). ‘Empirical Investment Equations for Developing Countries’, inLuis Serven and Andrés Solimano (eds),Striving for Growth After Adjustment: TheRole of Capital Formation. Washington, DC: World Bank, 107-43.
Sachs, Jeffrey (1989). ‘The Debt Overhang of Developing Countries’, in Jorge deMacedo and Ronald Findlay (eds),Debt, Growth and Stabilisation: Essays inMemory of Carlos Dias Alejandro. Oxford: Blackwell, 80-102.
Sachs, Jeffrey, Kwesi Botchwey, Maciej Cuchra, and Sara Sievers (1999).‘Implementing Debt Relief for the HIPCs’ Cambridge: Center for InternationalDevelopment, Harvard University. Available at: www.cid.harvard.edu/ .
Senhadji, Abdelhak (2000). ‘Sources of Economic Growth: An Extensive GrowthAccounting Exercise’.IMF Staff Papers, 47 (1): 129-57.
United Nations Conference on Trade and Development (UNCTAD) (2000).The LeastDeveloped Countries Report. New York: United Nations. Partly available at:www.unctad.org/en/pub/ps1ldc00.en.htm .
United Nations Environment Programme (UNEP) and World MeterologicalOrganization (WMO), Intergovernmental Panel on Climate Change (IPCC) (2001).Third Assessment Report—Climate Change 2001. Summaries available at:www.ipcc.ch/ .
United States General Accounting Office (GAO) (2000).Developing Countries: DebtRelief Initiative for Poor Countries Facing Challenges. Washington, DC:United States General Accounting Office. Available at: www.gao.gov.new.items/ns00161.pdf .
33
World Bank (2000).Intensifying Action against HIV/AIDS in Africa: Responding to aDevelopment Crisis.Washington, DC: World Bank, Africa Region.
World Bank (2001a).Global Development Finance. Washington, DC: World Bank.
World Bank (2001b). ‘Financial Impact of the HIPC Initiative’. Paper produced by theHIPC Unit. Washington, DC: World Bank. Available at: www.worldbank.org/hipc .
Young, Alwyn (1995). ‘The Tyranny of Numbers: Confronting the Statistical Realitiesof the East Asian Growth Experience’.Quarterly Journal of Economics,110 (3) 641-80.
